Gene- and evidence-based candidate gene selection for schizophrenia and gene feature analysis
Received 16 August 2008; received in revised form 7 July 2009; accepted 20 July 2009.
Abstract
Objective
Schizophrenia is a chronic psychiatric disorder that affects about 1% of the population globally. A tremendous amount of effort has been expended in the past decade, including more than 2400 association studies, to identify genes influencing susceptibility to the disorder. However, few genes or markers have been reliably replicated. The wealth of this information calls for an integration of gene association data, evidence-based gene ranking, and follow-up replication in large sample. The objective of this study is to develop and evaluate evidence-based gene ranking methods and to examine the features of top-ranking candidate genes for schizophrenia.
Methods
We proposed a gene-based approach for selecting and prioritizing candidate genes by combining odds ratios (ORs) of multiple markers in each association study and then combining ORs in multiple studies of a gene. We named it combination–combination OR method (CCOR). CCOR is similar to our recently published method, which first selects the largest OR of the markers in each study and then combines these ORs in multiple studies (i.e., selection–combination OR method, SCOR), but differs in selecting representative OR in each study. Features of top-ranking genes were examined by Gene Ontology terms and gene expression in tissues.
Results
Our evaluation suggested that the SCOR method overall outperforms the CCOR method. Using the SCOR, a list of 75 top-ranking genes was selected for schizophrenia candidate genes (SZGenes). We found that SZGenes had strong correlation with neuro-related functional terms and were highly expressed in brain-related tissues.
Conclusion
The scientific landscape for schizophrenia genetics and other complex disease studies is expected to change dramatically in the next a few years, thus, the gene-based combined OR method is useful in candidate gene selection for follow-up association studies and in further artificial intelligence in medicine. This method for prioritization of candidate genes can be applied to other complex diseases such as depression, anxiety, nicotine dependence, alcohol dependence, and cardiovascular diseases.
aDepartment of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN, USA
bDepartment of Psychiatry, Vanderbilt University Medical Center, Nashville, TN, USA
cDepartment of Cancer Biology, Vanderbilt-Ingram Cancer Center, Nashville, TN, USA
Corresponding author at: Vanderbilt University Medical Center, Department of Biomedical Informatics, 2525 West End Avenue, Suite 600, Nashville, TN 37203, USA. Tel.: +1 615 343 9158; fax: +1 615 936 8545.
ABSTRACT